How does the optimizer choose a set of parameters

[HelloHello]

When doing a walk forward optimizer, how does the optimizer choose a parameter set on which to optimize, the highest value right?

[NinjaTrader_Dierk]

Not sure I follow. The optimizer tests all possible parameter combinations.

[HelloHello]

Thank you for your reply.

But, when it does a test on out of sample data, it has to take a combination of parameters from the insample data part to test. How does it chose which set of parameters to use?

Like say the in sample data from 2005 to 2006 produces the following list of parameters that are the top 4 results. On the out of sample data test it has use a parameter combination. Which combination would it use, 16/15 because it is the highest? or 4/8 because it is the 2nd highest and the neighboring parameters have good results too?

16 15 $1011
4 8 $484
5 8 $364
7 7 $301

[NinjaTrader_Dierk]

Are you doing a walk-forward test or an optimizer run?
- the former takes the parameter combination which returned be best results of the previous optimizer run according the selected optimizer target ("Optimize on" parameter)
- the latter does not know about out-of-sample or in-sample logic